The Air Products Vattenfall Oxyfuel CO2 Compression - Science Direct

Available online at www.sciencedirect.com

Energy Procedia 37 (2013) 1490 – 1499

GHGT-11

The Air Products Vattenfall Oxyfuel CO2 Compression and Purification Pilot Plant at Schwarze Pumpe Vince White a*, Andrew Wrighta, Stephanie Tappeb and Jinying Yanc a

Air Products PLC, Hersham Place, Molesey Road, Walton-on-Thames, Surrey, KT12 4RZ, UK b Vattenfall Europe Generation AG, Hermann-Löns-Straße 3, 03050 Cottbus, Germany c Vattenfall Research & Development AB, 162 87 Stockholm, Sweden

Abstract Air Products has in the last couple of years been working with Vattenfall to realise the ambition of building the 2 from a commercial-scale coal-fired oxyfuel boiler. The pilot demonstrates three concepts that will together lead to an optimised, lowest cost, highest efficiency oxyfuel CO2 capture plant: sour compression, auto-refrigeration and membrane recovery of CO2 and O2 from the vent stream. This pilot plant will allow important data to be collected to assist in the design and optimisation of the next step: a demonstration plant. The technology supplied by Air Products was one of several technologies being evaluated by Vattenfall for the oxyfuel demonstration project that had been proposed in Jänschwalde, Germany. The agreed test program between Vattenfall and Air Products contains a number of identified R&D topics. The main results from the first tests will be presented in this paper. These initial tests involve establishing material and energy balances over the different main plant components as a first step. This will be continued by parametric testing on the warm-end of the pilot plant over the flue gas cooler/condenser and acid removal columns with the main goal of understanding the influence of these parameters on the SO2 and NOx chemistry and support process modelling activities.

© byby Elsevier Ltd.Ltd. © 2013 2013The TheAuthors. Authors.Published Published Elsevier Selection responsibility of GHGT Selectionand/or and/orpeer-review peer-reviewunder under responsibility of GHGT

Keywords: Type your keywords here, separated by semicolons ; Oxyfuel ; CO2 Purification; CPU; SOx; NOx; sour compression

* Corresponding author. Tel.: +44 1932 249948. E-mail address: [email protected]

1876-6102 © 2013 The Authors. Published by Elsevier Ltd.

Selection and/or peer-review under responsibility of GHGT doi:10.1016/j.egypro.2013.06.024

1491

Vince White et al. / Energy Procedia 37 (2013) 1490 – 1499

1. Introduction Air Products has in the last couple of years been working with Vattenfall to realise the ambition of 2 from a commercial-scale coal-fired f oxyfuel boiler. This pilot demonstrates three concepts that have been in development for a number of years: sour compression, auto-refrigeration and membrane recovery of CO2 and O2 from the vent stream. Together these technologies lead to an optimised, lowest cost, highest efficiency oxyfuel CO2 capture plant. The CO2 compression and purification pilot plant is connected via a slipstream to th oxyfuel pilot plant in Schwarze Pumpe, Germany. The slip stream makes up about 3% of the total flue gas flow rate, equivalent to around 1 MWth. The collaboration on the construction of the pilot was started in 2008. The CO2 compression and purification slip stream pilot plant is owned and operated by Vattenfall. All equipment has been delivered by Air Products and the plant was commissioned and handed over to Vattenfall in late 2011. In this paper, Air Products together with Vattenfall will present the initial results of the experimental program being carried out under a joint collaboration agreement between the two companies. An initial phase of experimentation was carried out and reported on at the 2 nd Oxyfuel Conference in 2011 [1]. In this paper further results are presented from extended testing that took place in the first half of 2012. 2. The Air Products Vattenfall Oxyfuel CO2 Compression and Purification Pilot Plant A simplified flowsheet of the Air Products Vattenfall Oxyfuel CO2 Compression and Purification Pilot Plant, known as the ACPP, is shown in Figure 1. This flowsheet shows the main parts of the th oxyfuel pilot plant, known as the OxPP. A slipstream of raw flue gas is taken from the OxPP and fed to the ACPP into a direct contact cooler where the water in the flue gas is condensed and the flue gas cooled before being fed to the first stage of compression. In order to control the amount of SO 2 being fed to the ACPP the slipstream from the OxPP ation unit (FGD). An intermediate concentration of SO2 can be achieved by drawing from both locations. CO2 Returned To OxPP Warm End Cold End

Flue Gas

Cooler & Condenser

Sour Compression

Condensate Collection

Condensate Collection

Process Condensate

Process Condensate

TSA & Mercury Removal

Auto-Refrigerated Inerts +O2 Removal Process

Air Products PRISM® Membrane CO2/O2 Recovery O2 and CO 2 Rich [To OxPP]

Inerts Vent [To OxPP]

Fig. 1. Overall flowsheet of the CO2 Pilot Plant showing where it connects to the main pilot plant, the OxPP

1492


2.1. Sour Compression One of the most interesting parts of the Air Products pilot plant is the process by which the raw CO 2 is compressed and purified using a concept that Air Products calls the sour compression process, which removes SOx and NOx from the feed CO2 during compression. The theory of this process, first described in a paper by Air Products at the GHGT8 conference [2], is that during compression to 30 bar most of the SOx and NOx in the raw CO2 will react to form acids. Acid formation could have a negative impact if this happens unexpectedly, or in the wrong place, but since Air Products discovered this concept in 2005 these sour compression reactions have been utilised in a configuration that can be used to remove SO x and NOx from the raw CO2 stream by controlling the formation of acids, potentially saving expensive upstream control options and minimizing potential downstream corrosion. Air Products has developed the sour compression process through lab scale experiments with Imperial College and more recently as part of a DOE cooperative agreement and a new purpose-built lab. The first experimental results confirming the theory were presented at GHGT9 [3]. Subsequent publications [4,5,6,7] presented results of experimental work to refine the understanding of the process. The results from the pilot plant at Schwarze Pumpe will be important as this process is now scaled to the demonstration plant scale, and we learn how the process is best optimised to work with or in place of conventional flue gas cleaning technologies. Figure 2 shows a flowsheet of the sour compression part of the ACPP . The slipstream from the OxPP enters a flue gas cooler condenser (FGCC) to cool the flue gas and allow water to be condensed from the stream. A 2-stage diaphragm compressor with an interstage cooler and condensate separator is then used to raise the pressure of the flue gas to 15 bar, where it enters the first column. This 15 bar column is a packed column with condensate from the bottom being cooled and recirculated to the top of the column, to allow residence time and vapour liquid contacting for the sour compression reactions to occur. Additionally there is a small fresh water makeup stream entering the top of the column. The flue gas stream from the top of the column is then compressed in a single stage diaphragm compressor to 30 bar, before entering the second column also arranged with condensate recirculation and makeup water addition. On leaving the top of the 30 bar column the stream is dried in a Temperature Swing Adsorption unit (TSA) and passed to the low temperature part of the process. 15 bar compressor

30 bar compressor 30 bar CO2 to TSA Make-up water

FGCC

cw

Fig. 2. Sour Compression to 30 bar


2.2. Auto-Refrigerated CO2 Purification The core of the ACPP is the low temperature partial condensation process used to purify the raw CO 2 once it has been compressed and dried. The purpose of this process is to remove the atmospheric gases nitrogen, oxygen and argon from the CO2 all of the feed gas will condense as the stream is cooled to near CO2 solidification temperatures. The key to the efficiency of this low temperature process is the refrigerant. Rather than using a conventional ammonia cycle, here the CO2 product itself is used to cool the raw CO2, making this the first oxyfuel CO2 purification plant to use this auto-refrigerated concept, which will deliver significant power reduction savings when applied to commercial scale plants. During the partial condensation process the argon, nitrogen and oxygen in the feed stream will mostly stay in the vapour phase and be separated as an inert vent stream. Some though will remain dissolved in the CO2, but these impurities can be removed using a distillation column integrated into the partial condensation process, producing a CO2 product stream with greater than 99% CO2 purity. One of the areas of study will be to monitor the performance of this process. Initial results of the operation of this part of the process have previously been presented [1] where it was discussed that the disturbances from the variation in the composition of CO2 in the flue gas from the OxPP needs to be expected and mitigated. Low pressure and medium pressure CO2 is evaporating in the main exchanger to provide refrigeration, cooling and condensing the feed stream of flue gas, but the temperature at which the CO2 starts to condense is affected by the composition and the pressure of that feed stream. The ACPP has the ability to change the feed pressure, but the composition is dependent on the combustion process. So the pressures of the evaporating MP and LP CO2 need to be controlled to maintain the exchanger temperature profile. When high purity CO2 is required and the distillation column is in operation, this provides another level of integration at the column reboiler, and this requires particular attention to maintain the correct reboiler temperature difference driving force. Our experience during commissioning meant that we adopted a feed forward control system that altered the pressures in the system depending upon the feed CO2 composition. 2.3. CO2 and O2 Recovery from the Vent Stream The final process step in the pilot plant addresses the inerts stream that leaves the low temperature process, and how to maximise recovery from this stream which would otherwise be vented to the atmosphere. The composition of this stream is determined by the vapour phase equilibrium at the cold end of the process, and so this stream will contain CO2. The amount of CO2 leaving through this vent stream determines how much CO2 would be lost in this process. What is needed is a way to recover the CO2 in this inerts stream as well Products has installed a series of PRISM® membrane modules that will demonstrate the recovery of CO2 from the vent stream so that on a commercial plant this CO2 can be recovered back to the process. An additional bonus is that whilst the membrane is recovering CO 2 it is also recovering oxygen. When this oxygen is returned to the boiler with the co-recovered CO2, this reduces the amount of oxygen required from the air separation unit, making the ASU smaller and reducing the power required.

1493

1494


3. Experimental Programme The experimental programme has been separated into five phases, shown in Table 1. The purpose of Phase 1 was to ensure that all the equipment worked, lessons learnt were documented and initial results were reported. This work took place as part of the commissioning of the ACPP. The primary goal of Phase 2 was to obtain all data required to complete a mass and energy balance on each individual unit at a non-constrained design point. This work has been completed. Phase 3 is a set of 2/NOx removal columns. Gas flows, liquid flows, feed SO 2 composition, temperatures and pressures will be varied around the system to map out the SO2/NOx/Hg removal at a variety of operating conditions. The results presented in this paper are from this parametric testing phase. Table 1. Testing Phases as Defined in the R&D Agreement Phase

Weeks

1

Start-up and Operating Envelope

Commissioning

2

Baseline Tests

2

3

Parametric Testing

8

4

Cold End

24

5

Warm End including Reactors, Membranes, TSA and Guard Bed

24

-purity CO2 In P and more complicated cycles including a stripping column for enhanced O 2 removal. The final phase of arm, evaluate the performance of the TSA adsorbent 3.1. Operational Constraints th oxyfuel testing facility, the OxPP, has to be running for the ACPP to operate. Testing on new burner developments takes place on the OxPP, which places operational constraints on the ACPP. Several burners have been tested at the OxPP and each requires commissioning time, reducing the time available for running the ACPP. Each has different operational characteristics that, for instance, vary the NOx in the feed to the ACPP, affecting its performance, and the excess oxygen, affecting the raw CO2 purity which affects the performance of the cold end of the process.

3.2. Analysis The list of analysers that are currently being used in the ACPP are shown in Table 2. Analyser 1, the SICK analyser, has the difficult job of measuring SO2, NO and NO2 in the low pressure streams with significant water content. The ACPP and OxPP are both using this new type of analyser from SICK Maihak. Each analyser is unique and requires its own multi-dimensional calibration curve to compensate for NO2, H2O, CO2 (etc) cross interference. Results from the SICK analysers have been cross-checked with Teledyne and Rosemount analysers where possible.

1495


It also has to be appreciated that the chemistry that converts SOx and NOx to acid streams will also occur in the analyser lines, making it a challenge to be able to differentiate the composition at the analyser from the composition in the process stream, the latter being the parameter we are interested in, but the former being the value that the analyser will show. Minimising this effect is important and was addressed by proper design of the analyser lines. Lab experiments indicated that N2O could form under some conditions and so the Horiba N2O analyser was added to allow us to understand when this could occur. The analysers were all connected to several streams to allow them to sample from different locations. This therefore means that not all of the streams can be analysed at the same time. Liquid analysis for nitrate, nitrite, sulphate and sulphite is also being carried out on the condensate from several parts of the plant to allow cross-checking with the mass balance derived from gas phase analysis. Table 2. Analysers used in testing

1

Manufacturer SICK

Model MCS100E H/W

2 3 4 5 6

Teledyne Teledyne Teledyne Teledyne Teledyne

6400EH 9110EH 7500 3000PA 4060

7 8

Teledyne Teledyne

3000TA 4060

9

Horiba

VIA-510

Measurement Type Infrared filter photometer with a zirconia oxide sensor UV Fluorescence Chemiluminescence Non-dispersive infrared Electrochemical (fuel cell) Gas chromatography with thermal conductivity detector Electrochemical (fuel cell) Gas chromatography with thermal conductivity detector Non-dispersive infrared

Components ppm HCl, ppm CO, ppm SO2, ppm NO, ppm NO2, %CO2, %H2O, %O2 ppm SO2 ppm NO, ppm NO2 %CO2 %O2 %N2, %O2+Ar ppm O2 ppm N2, ppm O2+Ar ppm N2O

4. Phase 1 and 2 Results - Sour Compression Results were presented previously showing the first experience of SO 2 and NOx removal in the ACPP [1]. These results, shown in Figure 3, showed significant SO2 removal across the sour compression process in the ACPP. Note that SO2 drops 10-20% across the flue gas cooler condenser (FGCC) and that part of the sour compression process happens within the 15 bar compressor, as it consists of two stages of diaphragm compression, with interstage cooling and condensate knockout, which provides enough residence time for a significant amount of reaction of SO2 and NOx to occur. Also NO is seen to convert to NO2 during compression, as predicted by the sour compression process. 600

3000

400

NO

Inlet Compressor

Inlet FGCC

Inlet 15 bar Column

SO2 (ppm Dry)

NOx (ppm, Dry)

Inlet FGCC

Inlet Compressor

2000

Exit 15 bar Column

200

Inlet 15 bar Column

1000

NO2 0

Exit 15 bar Column 0

a) NO converts to NO2 during compression as predicted by the sour compression process

b) SO2 reduces during compression

Fig. 3. NOx and SO2 removal over the first stage of the sour compression process

1496


5. Phase 3 Results - Parametric Testing of the Sour Compression Process armPhase 3 testing started in May 2012 and consists of pilot plant, i.e. the flue gas cooler condenser and SO2/NOx removal columns. Gas flows, liquid flows, feed SO2 concentration, temperatures and pressures are varied around the system to map out the system performance at a variety of operating conditions. The results of a few of these tests are presented and discussed below.

80

80

SO2 Removal from Flue Gas (%)

100

SO2 Removal from Flue Gas (%)

100

Equal NOx 60

40

20

0 0

2

4

6

8

60

40

20

0

10

9

10

11

12

13

14

15

Feed SO2/NOx (ppm/ppm)

Pressure (bara)

a) SO2 removal from inlet of the 15 bar compressor to the outlet of the 15 bar column as SO2/NOx varied. Diamonds indicate a variation in SO2 with NOx constant and squares indicate NOx has varied with SO2 constant

b) SO2 removal from inlet of the 15 bar compressor to the outlet of the 15 bar column as compressor outlet/column pressure is varied 1.20

Relative SO2 Removal across 15 bar column (-)

Relateive SO2 Removal across 15 bar column (-)

1.20

1.00

1.00

0.80

0.80

0.60

0.60

0.40

0.40

0.20

0.20

0.00

0.00 0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Relative Residence Time in 15 bar column (-)

Relative Fresh Make-up Water Flow to 15 bar column (-)

c) SO2 removal from inlet to outlet of the 15 column as flowrate, hence residence time, is varied. Results have been normalised

d) SO2 removal from inlet to outlet of the 15 column as makeup water to the top of the 15 bar column is varied. Results have been normalised

Fig. 4. Sensitivities in SO2 removal


The results shown in Figure 4 represent the results of four sensitivity experiments that will now be described separately. 5.1. SO2 Removal as a Function of SO2/NOx ratio The SO2 concentration of the flue gas from the OxPP to the ACPP can be varied by taking flue gas from before the FGD, after the FGD, or a mixture of both. NO x to the ACPP cannot be controlled, but it did vary between tests, from around 260 360 ppm. By examining the data where other important parameters have been held constant it is possible to construct the plot shown in Figure 4 a) which shows how the SO2 removal from the flue gas changes as a function of the ratio of SO2 to NO + NO2 (ie NOx) in the feed. In this case the inlet condition is after the flue gas cooler condenser (over which there is a 1020% removal of the SO2 from the flue gas) and before the 15 bar compressor, so the removal indicated in Figure 4 a) includes SO2 removed during compression and in the 15 bar column. Further conversion of course happens in the 30 bar column. One can see from this plot how SO2 removal reduces with increasing SO2/NOx ratio for this fixed piece of equipment. The diamonds in this plot indicate where NOx has remained constant but the SO2 content in the flue gas has increased and led to reduced SO2 removal whereas the squares show where NO x has reduced with a constant SO2 content in the flue gas, leading to a reduction in SO2 removal. These results indicate the importance of NOx in the flue gas to the removal of SO2 from the flue gas. An over-enthusiasm for NOx reduction in combustion is detrimental to the sour compression process, as has long been predicted by experimental work and modelling studies. 5.2. SO2 Removal as a Function of Pressure Pressure is a parameter that has long been predicted as being important for the sour compression process, predicted by the kinetics of the reactions involved. This sensitivity is clearly shown in Figure 4 b) where the SO2 removal is shown at two different pressures. As in Figure 4 a) the conditions are taken from after the FGCC and to the outlet of the 15 bar column. All other parameters were held constant and the pressure in the 15 bar column reduced to 11 bara. SO2 removal in the compressor and column reduce with lower pressure. 5.3. SO2 Removal as a Function of Column Residence Time In order to determine the effect residence time in the 15 bar column has on conversion, the flow to the column was reduced. Since the 15 bar compressor is a positive displacement machine, changing the flowrate can only be achieved by opening the recycle valve and allowing flow to recycle around the compressor. Doing this means we can produce the plot shown in Figure 4 c). The residence time in the column increases as the flow through the compressor is reduced and this leads to an increase in SO2 conversion. The results shown in Figure 4 c) are for the column alone: the inlet condition is the feed composition to the 15 bar column, not the composition from the FGCC. The results have also been normalised and from this one can see that, roughly doubling the residence time doubles removal of SO2 across the 15 bar column.

1497

1498


5.4. SO2 Removal as a Function of Column Make-up Water Flowrate The final parametric variation to report is the sensitivity of SO2 removal to the flowrate of fresh makeup water to the top of the 15 bar column. As can be seen in the normalised results of SO 2 removal across the 15 bar column in Figure 4 d) which were carried out with high SO2 in the feed, around 2000 ppm, there is a very strong relationship between these parameters. Water is a reactant in the sour compression process and, with this high sulphur case, increasing the amount of water fed to the column improves conversion. Note that SO2 will to some extent dissolve in the extra added water, thereby increasing SO2 removal, but not by the sour compression process. Analysis of the condensate formed is therefore important to determine whether sulphite or sulphate is being produced. 5.5. NOx Removal In addition to the sensitivity results shown in Figure 4 it should be noted that the results of NOx removal up to the exit of the 15 bar column indicate that in most cases around 80% of the NO+NO 2 was removed from the gas phase at this point, with the rest being eliminated in the 30 bar column. It was found, however, that in certain situations, particularly high SO2/NOx ratios, some of the NOx would convert into N2O. In the worst case, all of the NOx converted into N2O. Further work is required to understand this behaviour. 6. Conclusions This pilot plant has allowed important data to be collected that will assist in the design and optimisation of any future demonstration plant. The agreed test program between Vattenfall and Air Products contains a number of identified R&D topics. The main results from the first tests have been presented in this paper. This testing will be continued in the autumn with further parametric testing on the warm-end of the pilot plant when the plant restarts with the main goals of understanding the influence of these parameters on the SO2 and NOx chemistry and supporting process modelling activities. The Air Products Vattenfall Oxyfuel CO2 Compression and Purification Pilot Plant at Schwarze Pumpe is the first demonstration of sour compression in representative equipment and the first demonstration of auto-refrigerated inerts removal. Although there has only been limited operation so far, results are very encouraging and support the sour compression theory.


References [1] -Air Products Oxyfuel CO2 Compression and Purification Pilot Plant at Schwarze Pumpe 16th September 2011, Capricorn Resort, Yeppoon, Queensland, Australia [2] -Derived CO2 for Sequ Conference on Greenhouse Gas Control Technologies GHGT-8 Conference Proceedings, Trondheim, Norway, 2006. [3] White, V., Torrente-derived CO2 presented at the 9th International Conference on Greenhouse Gas Control Technologies, GHGT-9; Washington DC; 16th 20th November 2008, Energy Procedia 1 (1), pp. 399-406, 2009 [4] Torrentex and NOx from Oxyfuel-Derived CO2 Combustion Conference, 8th 11th September 2009, Radisson Hotel, Cottbus, Germany [5] White, V., Torrente-Derived CO2 International Journal of Greenhouse Gas Control, 4(2), 137-142. [6] TorrenteSOx and NOx from oxyfuel-derived CO2 nce on Greenhouse Gas Control Technologies GHGT-10; Amsterdam; 19th 23rd September 2010 Energy Procedia, 4, 908-916 [7] TorrenteCompression Process for the Purification of Oxyfuel-Derived CO2 1), S224-S230

1499